1. Giussani

Antiepileptic drugs discontinuation by people with epilepsy in the

general population

Giorgia Giussani BiolD1, Elisa Bianchi MSc1, Valentina Canelli PharmD1, Giuseppe Erba MD2,

Carlotta Franchi PharmD1, Alessandro Nobili MD1, J. W. Sander FRCP3,4, Ettore Beghi MD1, and

the EPIRES Group*

1 IRCCS-Istituto di Ricerche Farmacologiche “Mario Negri”, Milano, Italy.

2 Department of Neurology, SEC, University of Rochester, Rochester, NY, United States.

3 Stichting Epilepsie Instellingen Nederland (SEIN), Heemstede, 2103 SW, The Netherlands.

4 NIHR University College London Hospitals Biomedical Research Centre, UCL Institute of

Neurology, Queen Square, London, WC1N 3BG, and Chalfont Centre for Epilepsy, Chalfont St

Peter SL9 0RJ, UK.

* EPIRES Group = Neurologists and child neurologists: Elio Agostoni, Larissa Airoldi, Francesco

Basso, Marialuisa Carpanelli, Mariolina Di Stefano, Andrea Magnoni, Ottaviano Martinelli, Andrea

Rigamonti, Andrea Salmaggi, Lorenzo Stanzani, Cristina Volpe, Nicoletta Zanotta, Claudio Zucca.

General practitioners: Marco Agudio, Flavio Arienti, Graziano Rocco Arrigoni, Giancarlo Balestra,

Piercarla Ballabio, Stefano Marco Balossi, Maria Rita Baratti, Fabrizia Enrica Barteselli, Fausto

Bellani, Mauro Bellani, Gianpiero Bellini, Pier Aldo Beretta, Roberto Beretta, Massimo Bergamini,

Annalisa Bertella, Francesco Bertolini, Rolando Bevilacqua, Marcellino Arturo Bianchi, Patrizia

Bigiolli, Giuseppe Binda, Gigliola Biondelli, Libero Bodega, Edoardo Giovanni Bolis, Marco

Vincenzo Carlo Bolla, Carlo Ambrogio Bonacina, Marco Bonfanti, Fausta Borghetti, Roberto

Brambilla, Wilma Brancaleone, Francesco Brusadelli, Sabine Bub, Giovanna Caccia, Myriam

Caglio, Carlo Calvi Rossi, Daniela Capra, Davide Carone, Giuseppe Carì, Silvio Carrera, Rosalia

Cavenago, Maria Luisa Cecchetti, Bice Antonella Centonze, Fabio Cereghini, Ciro Cerrone, Maria

2. Giussani

Pia Ciappetta, Bruno Cogliati, Ombretta Colombano, Alberto Colombo, Daniela Colombo, Maria

Lia Corti, Nicla Cremonini, Marco Crippa, Rolando Antonio Crippa, Mario Crotta, Salvatore Curto,

Patrizia Daielli, Domenico De Gilio, Giuseppe Pasquale De Pascalis, Elena Decet, Luigi Della

Morte, Felice Duvia, Antonia Favorito, Roberto Ferrario, Luigi Fezzi, Angela Laura Finzi, Bruno

Ugo Fiorentino, Agnese Fornaciari, Silvana Franceschetti, Fabio Fulconis, Giovanni Fumagalli,

Maurizio Fumagalli, Giuseppe Galbiati, Carlo Andrea Galimberti, Donatella Gecchele, Barbara

Ghiazza, Paolo Ghislanzoni, Laura Gioffredi, Taraneh Hassibi, Simone Julita, Giuseppe Leone,

Claudio Levi, Giovanni Locatelli, Anna Alessandra Locati, Enrico Giovanni Longhi, Ingrid

Lukacova, Mauro Giovanni Manfroi, Paola Manzoni, Claudio Marcolini, Fabio Minicucci,

Mohammed Safwan Martini, Rita Masperi, Sussele Mayan, Valeria Mazzoleni, Antonio Menga,

Fiorella Merlini, Enrico Messina, Marta Micheli, Rocco Micò, Andrea Millul, Barbara Montini,

Alberto Palazzuolo, Silvano Primo Panzeri, Giovanni Passoni, Laura Pensotti, Maria Marcella

Petrone, Roberto Pezzuto, Elisabetta Pigazzini, Nicoletta Pirovano, Margherita Pontiggia, Roberto

Pozzi, Carlo Quadrelli, Filomena Rampello, Marco Realini, Giovanni Righetti, Cesare Ripa, Sonia

Rivetta, Antonino Romeo, Carlo Rota, Tiziana Rusconi, Andrea Santarpia, Ivana Santi, Luigi

Sciani, Gianni Scoccimarro, Eugenio Scopinaro, Angelo Giuseppe Semerano, Daniela Sferco, Anna

Ersilia Sirtori, Angela Spandri, Bruno Spinelli, Fausto Stefanoni, Raffaele Taiana, Libero

Tamagnini, Laura Tassi, Giorgio Tentori, Augusto Teodoro, Katerina Tinterova, Ampelio Tocchetti,

Michela Todeschini, Lorenzo Trezza, Valter Valsecchi, Roberto Vanini, Claudio Felice Vercelloni,

Anna Villella, Maria Viola, Pietro Vismara, Daniele Mario Zoboli.

Corresponding author: Ettore Beghi, MD

Laboratory of Neurological Disorders

IRCCS-Mario Negri Institute for Pharmacological Research

Via Giuseppe La Masa 19,

20156 Milan, Italy.

Tel. +39 02 39014542 fax +39 02 39001916

3. Giussani

e-mail: ettore.beghi@marionegri.it

Running title: Antiepileptic drugs discontinuation

Keywords: Epilepsy, antiepileptic drugs, discontinuation

Number of text pages: 21

Number of words: 3368

Number of references: 41

Number of figures: 1

Number of tables: 4

Supporting information: Supporting Table S1 and Supporting Table S2

4. Giussani

Abstract

Objective: Rate, reasons and predictors of antiepileptic drugs (AEDs) discontinuation were

investigated in a well-defined cohort of people with epilepsy to verify efficacy and tolerability of

treatment in up to twenty years from treatment initiation.

Methods: The history of AED usage in children and adults with epilepsy registered with 123 family

physicians in an area of Northern Italy between 2000 and 2008 was recorded. Cumulative

probabilities of AED withdrawal for specific reasons were estimated using cumulative incidence

functions. The probabilities of withdrawing for terminal remission, or of achieving sustained

remission while still on treatment, were also evaluated. The roles of sex, age at diagnosis, seizure

types, duration at diagnosis and syndrome were assessed with hazard ratios (HR) and 95%

confidence intervals (95% CI).

Results: 731/747 individuals were treated with one or more AEDs during the disease course. The

three commonest drugs were valproate, carbamazepine and phenobarbital. Reported reasons for

AED withdrawal were, in decreasing order, terminal remission, ineffectiveness and adverse events.

The probability of withdrawing the first AED for terminal remission was 1.0% at one year and

increased to 20.0% at twenty years. Corresponding rates for ineffectiveness were 2.9% and 12.6%;

and for adverse events 0.5% and 3.3 %. Reasons for withdrawal varied with individuals’ age, sex,

disease characteristics, and drugs.

Significance: The initial AED given was retained in the majority of cases. Terminal remission, lack

of efficacy and adverse effects were, in decreasing order, the commonest reasons for AED

discontinuation. Withdrawal could be predicted by age at diagnosis, sex and clinical characteristics

and varies among drugs.

Keywords: Epilepsy, antiepileptic drugs, discontinuation

5. Giussani

Introduction

The response to antiepileptic drugs (AEDs) is an important indicator of the prognosis of epilepsy.

AEDs are generally efficacious in controlling seizures. In randomized trials, but also in clinical

practice, seizure freedom is seen in about 50% of people with previously untreatedepilepsy.1-3 In

people who do not respond to the first AED, the use of a second drug leads to complete seizure

control in a lower proportion of cases and there is a further decrease with each subsequent treatment

change.4-7 If the first drug fails, however, seizure-free rates vary across studies and, although drug-

resistant epilepsy has been defined as the failure of two appropriate AEDs,8 there are reports of

people achieving seizure freedom even after having failed several AEDs.5, 9-11 While withdrawal of

ineffective drugs, when given as first or subsequent treatments, is usually clearly documented, the

information becomes fragmentary regarding the reasons for discontinuation. Additionally, with

some exceptions,12-14 AED retention has not been assessed in a population-based sample. Lastly,

there are no data on the history of AED treatment in samples of children and adults with epilepsy

from the same population and followed for a prolonged period of time.

We investigated rate, reasons and predictors of AED discontinuations in a well-defined population

followed for up to twenty years from treatment start. We aimed: 1. To estimate the retention rate of

AEDs according to the order of administration (first, second, or subsequent) in general and by drug;

2. To investigate the reasons for stopping AEDs with reference to the sequence of drug assignment;

3. To ascertain possible predictors of treatment discontinuation with reference to the commonest

reasons for discontinuing the assigned drug.

Methods

The study sample (146,506; year 2008), representative of the general population, was based in the

province of Lecco, Northern Italy.15,16

6. Giussani

The details of the study design are given elsewhere15,16 and are summarized here. In Italy, the

diagnosis of epilepsy is confirmed by a neurologist who also takes the responsibility to start and,

where needed, to change or stop the treatment. The individual is then followed by his/her general

practitioner (GP) to obtain drugs free-of-charge. One hundred and twenty-three GPs (47% of those

active in the province) were invited to identify people with epilepsy in their practice’s list and who

had lived in the area for at least one year between January 1st 2000 and December 31st 2008. The

medical records of all these individuals in the practice and, if available, in other in- and out-patient

facilities such as hospitals, nursing homes, out-patient clinics were reviewed. Epilepsy diagnosis

was based on a history of two or more unprovoked seizures at least 24 hours apart17 and confirmed

by a neurologist. People with acute symptomatic seizures, neonatal seizures, single unprovoked

seizures, and paroxysmal non-epileptic events were excluded. For each eligible case, data were

collected on demographics (date of birth, sex), predominant seizure types,18 date of diagnosis,

epilepsy syndrome,19 number and type of drugs used since the beginning of treatment. For each

AED, details were collected on starting date, maximal maintenance daily dose, and withdrawal date

or last follow-up date, whichever came earlier. If the drug was discontinued, reasons for

discontinuation (ineffectiveness, adverse events, terminal remission or other) were recorded. As the

study period preceded the new epilepsies classification scheme,20 the 1989 classification19 was used.

As detailed information was not available in all cases, seizures and syndromes were classified

using broad categories. Seizures were classified as focal, generalized, or unclassifiable. Syndromes

were classified as partial (idiopathic, symptomatic, or cryptogenic), generalized (idiopathic or

symptomatic/cryptogenic), undetermined, and special. All data from the GPs and, if needed, from

the consulting neurologists (including those outside the study area) were de-identified and filed in a

central database by two of us (GG and VC) who interacted with the GPs and, if required, with the

neurologists. Two of us (EBe and GE) revised the information included in the database to confirm

the appropriateness of the indications of all drugs assigned to each individual during the disease

course.

7. Giussani

Prevalence and incidence of epilepsy, of drug-resistant epilepsy and prognostic patterns in the same

study population have been previously reported elsewhere.15,16

Statistical analyses were performed using SAS (version 9.2; SAS Institute, Inc, Cary, NC, USA).

Descriptive statistics are presented as counts and percentages. Administration frequencies and

cause-specific withdrawal frequencies were calculated for each active principle and by prescription

order. AEDs were also grouped in two different classes: old and newer (marketed before and after

1990). Old drugs included barbexaclone (BSC), carbamazepine (CBZ), clobazam (CLB),

clonazepam (CNP), ethosuximide (ESM), phenobarbital (PB), phenytoin (PHT), primidone (PRM),

valproate (VPA), valpromide (VPM); newer drugs included gabapentin (GBP), levetiracetam

(LEV), lamotrigine (LTG), oxcarbazepine (OXC), pregabalin (PGB), tiagabine (TGB), topiramate

(TPM), vigabatrin (VGB), zonisamide (ZNS).

To account for competing risks, cumulative probabilities of AED withdrawal from specific causes

over twenty years from treatment start were estimated using cumulative incidence functions. The

cumulative probability of cause-specific AED withdrawal was calculated for the first, second, and

third AED, for the most common AEDs (i.e. used by more than 100 people), and for new and old

AEDs separately. Differences in the cumulative incidence functions between new and old AEDs

were assessed using Gray’s test.21 For the first AED, the cumulative probability of withdrawing for

terminal remission (ie, seizure freedom for at least two years at last follow-up) while still on

treatment, was also evaluated.

The association between drug discontinuation and sex, age at diagnosis (<15, 15-64, 65+ years),

seizure types (focal, generalized, unclassifiable), duration from first seizure to diagnosis, and

epilepsy syndrome (idiopathic, cryptogenic/symptomatic, special/undetermined) was assessed using

the Cox proportional hazards function. Models for the most commonly used AEDs were adjusted

for the number of previous drugs taken. Results were presented as Hazard Ratios (HR) with 95%

Confidence Intervals (95%CI). Missing data were handled using the listwise deletion method.

8. Giussani

The study was approved by the Ethics Committee of the Provincial Hospital of Lecco (Protocol

2011-003428-11). Local Health Service authorization was obtained to collect de-identified data

from the GPs. In less than 5% of cases, the GPs needed to collect additional information from

individuals, after informed consent. All the data were managed according to the current Italian

privacy rules.

Results

The study sample consist of 747 people with epilepsy aged 11 months through 94 years and

followed for a total of 11,045.5 person-years (mean 14.8 years; interquartile range 4.5-22.5).

Clinical characteristics are provided in Table 1; 731 people (98%) were treated with at least one

AED.

The use of each compound as first, second, third, or fourth to ninth AED is shown in Table 2. The

three AEDs most commonly used as first drug were valproate, carbamazepine and phenobarbital

and the same drugs were also the commonest second option. The third option included, in

decreasing order, carbamazepine, levetiracetam and topiramate.

The commonest reasons for drug withdrawal were, in decreasing order, terminal remission,

ineffectiveness and adverse events. Table 3 shows the reasons for withdrawal with reference to the

sequence of drug assignment. For the first AED, the main reasons for withdrawal were terminal

remission followed by ineffectiveness; for the second AED the main reasons were ineffectiveness

followed by terminal remission; for the third and the fourth AED the main reason was

ineffectiveness, followed by adverse events. Withdrawal for ineffectiveness increased from the first

to the seventh drug, while withdrawal for terminal remission decreased progressively after the first

drug. Adverse events showed a moderate increase from the first to the last assigned drugs.

Reasons for drug withdrawal for each AED are provided in Table 4. The percentages of withdrawal

of carbamazepine, phenobarbital and valproate due to ineffectiveness were in general lower than the

percentages reported for other AEDs.

9. Giussani

Cumulative probabilities and predictors of withdrawal of the first, second and third AED

1. First AED. The cumulative probability of withdrawal of the first AED for ineffectiveness

increases from 2.9% at one year to 12.6% at twenty years (Supporting Table S1). The only predictor

was age at diagnosis: compared to those aged less than 15 years, those in the 15-64 year group were

less likely to withdraw due to ineffectiveness (HR 0.46, 95% CI 0.28-0.75), while those in the

oldest group showed no significant differences (HR 1.26, 95% CI 0.61-2.59). The cumulative

probability of withdrawing the first AED for terminal remission increased from 1.0% at one year to

20.0% at twenty years. The variables associated with first AED withdrawal due to seizure freedom

were age at diagnosis, sex and epilepsy syndrome. Compared to those aged less than 15 years, those

in the 15-64 year group were less likely to withdraw the drug (HR 0.58, 95% CI 0.40-0.86) while

those in the oldest group had a similar HR although this was not statistically significant. . Females

had a lower probability than males of withdrawing the drug (HR 0.56, 95% CI 0.39-0.82). Those

with cryptogenic/symptomatic epilepsies (HR 0.43, 95% CI 0.29-0.63) had a lower probability of

withdrawing the drug than those with idiopathic epilepsies. The cumulative probability of

withdrawing the first drug for adverse events was 0.5% at one year and increased to 3.3% at twenty

years. The probability of withdrawing the first drug for other reasons was 0.2% at one year and

6.6% at twenty years. Predictors for adverse events and other reasons were not assessed, due to the

small numbers. No significant differences were observed between old and new AEDs given as first

treatment (Supporting Table S2). Four hundred and sixty people never withdrew the first

antiepileptic treatment and 224 of them (50.9%) started a period of remission lasting until the end of

follow-up. The cumulative probability of either withdrawing the first AED for seizure freedom or of

achieving sustained remission while still on treatment was 23.1% at one year and increased to

48.0% at twenty years.

2. Second AED. The cumulative probability of withdrawing the second drug at twenty years for

ineffectiveness was 15.8%, for adverse events 4.3%, for terminal remission 13.3%, and for other

reasons 7.7% (Supporting Table S1). The only variable associated with discontinuation of the

10. Giussani

second AED for ineffectiveness was seizure type: compared to partial seizures, generalized seizures

were more likely to lead to drug withdrawal (HR 2.05, 95% CI 1.10-3.83). Due to the small number

of events, predictors for other reasons for the withdrawal of the second AED were not evaluated.

No significant differences were found between old and new AEDs (Supporting Table S2).

3. Third AED. The cumulative probability of withdrawing the third AED at twenty years for

ineffectiveness was 39.3%, for adverse events 8.0%, for seizure freedom 4.3%, and for other

reasons 5.0% (Supporting Table S1). Due to the small numbers, predictors for the third AED

withdrawal were not assessed. The comparison between old and new AEDs showed no significant

differences (Supporting Table S2).

Cumulative probabilities and predictors of withdrawal of the most commonly used AEDs

At 20 years, the cumulative time-dependent probability of withdrawal of carbamazepine for

ineffectiveness was 10.8%. The corresponding values were 13.5% for phenobarbital and 12.3% for

valproate (Figure 1A). The 20-year probability of withdrawal for terminal remission was 12.9% for

carbamazepine, 14.8% for phenobarbital and 27.4% for valproate (Figure 1B).

Predictors of withdrawal due to ineffectiveness were seizure type for carbamazepine and age for

phenobarbital. No predictors were found for valproate. Generalized seizures were more likely than

partial seizures to lead to withdrawal of carbamazepine for ineffectiveness (HR 3.07, 95% CI 1.35-

6.98). Individuals in the 15-64 years group had a lower probability of withdrawing phenobarbital

due to ineffectiveness than the youngest age group (HR 0.24, 95% CI 0.11-0.54), while those in the

oldest group showed no significant differences (HR 0.21, 95% CI 0.03-1.57).

Variables associated with drug withdrawal due to seizure freedom were sex for carbamazepine and

age, sex and syndrome for valproate. No predictors were found for phenobarbital. Females had a

lower probability of withdrawing carbamazepine for seizure freedom than males (HR 0.47, 95% CI

0.22-0.99). Compared to people in the <15 year group, those aged 15-64 years were less likely to

withdraw valproate due to ineffectiveness (HR 0.30, 95%CI 0.13- 0.70), while those in the oldest

11. Giussani

group showed no significant differences (HR 0.36, 95%CI 0.05-2.58); females had a lower

probability of withdrawing the drug than males (HR 0.52, 95% CI 0.30-0.91); compared with

idiopathic epilepsies, those with cryptogenic/symptomatic epilepsies were less likely to have drug

withdrawal (HR 0.35, 95% CI 0.19-0.65).

Due to the small number of events, cumulative probabilities and predictors of carbamazepine,

valproate and phenobarbital withdrawal for adverse events and other reasons were not assessed.

Discussion

People in our cohort, a representative sample of people with epilepsy in a defined area of Northern

Italy, have used a large number of AEDs over a 15-year period. However, during this time, the three

most common drugs given at the start of treatment were valproate, carbamazepine and

phenobarbital and only 10 percent of people started treatment with a new AED. Terminal remission

was the commonest explanation for discontinuation of the first drug (20% at twenty years),

followed by lack of efficacy (12.6%). Withdrawal of the first drug for adverse events was only

0.5% at one year and increased to 3.3% at twenty years. While the discontinuation of a drug for

terminal remission tended to decrease with AED order, treatment stop for ineffectiveness and for

adverse events tended to increase even though clear trends could not be detected because of the

small samples at the highest rankings. The reasons for drug withdrawal varied with age, sex and

disease characteristics.

The probability of retaining the first drug in the treatment schedule and starting a period of

remission lasting until the end of follow-up, or of stopping the first treatment for terminal

remission, was high, 48% at 20 years. Others found that the proportion of seizure-free individuals

on the first AED ranged from 5.4% to 62%;3,4,12,22-24 60.5% never withdrew the first AED and 51%

of them started a period of remission lasting until the end of follow-up. This finding supports the

concept that in clinical practice the majority of people with epilepsy can be easily controlled with

any of the available compounds even after long follow-up periods.

12. Giussani

In our study, the cumulative probability of discontinuing the first drug at 12 months for lack of

efficacy or adverse events was only 3.4%. Our findings are fairly similar to the results of a

Lebanese study of people with newly diagnosed focal epilepsy, which found a 12-month retention

rate of 93.6%.25 Our data only partly agree, however, with other long-term follow-up studies. In the

UK National General Practice Study of Epilepsy12 the first assigned drug was discontinued for lack

of efficacy in 21% of cases (compared to our 12.6%) but the discontinuation rate for adverse events

was 11.5% (compared to 3.3% in our study). This difference may be explained by the use of

carbamazepine, phenytoin, valproate or phenobarbital in 96% of the UK cases as compared to 87%

in our cohort. The use of fairly low daily doses for some drugs in our study (see Supporting Table

3) could be another explanation. No differences were found in retention rates when comparing old

and new AEDs. Our findings are in keeping with a study in children26 but differ from a study in

older adults27 in which the 12-month retention rates ranged from 12.5% (oxcarbazepine) to 90%

(valproate). In this latter study, however, the rates were calculated in people with refractory

epilepsy.

When comparing drugs in our study, differences in retention were seen. Discontinuation for lack of

efficacy was most common with GABAergic drugs while discontinuation for adverse events was

mostly seen with topiramate, phenytoin, carbamazepine and levetiracetam in decreasing order.

These findings are not unexpected even though the rates differ from those of other reports28-33 on

account of differing prescribing patterns and different populations at risk.

People in our cohort taking carbamazepine, valproate or phenobarbital discontinued the assigned

treatment for adverse events in 5.6, 1.1 and 2.8% of cases respectively. Our rates are significantly

lower than those reported by others (carbamazepine, valproate and lamotrigine stopped in 27, 13

and 10% of cases respectively).3 Possible explanations for this difference are the source population,

as the Scottish study was not population-based, and the use of different daily doses.

Children and elderly subjects tended to stop the first drug mostly for lack of efficacy and, less

frequently, due to terminal remission. Childhood and adolescent syndromes less responsive to the

13. Giussani

current treatments and the need to resort to complex therapeutic regimens in people with

comorbidities are possible explanations.

The lower rates of treatment withdrawal due to terminal remission in women than in men likely

reflects the higher proportion of females in the age class <15 years (59% versus 41% of male

children) who continued the first treatment, perhaps because of the fear of withdrawal in a period of

hormonal and emotional changes. As expected, people with idiopathic epilepsies and/or generalized

seizures were most likely to respond to the assigned treatments.

The study has strengths and limitations. The major strengths are the population base, the fairly large

sample size, and a long follow-up. In this regard, our findings apply to the general epilepsy

population in which, by definition, severe epilepsies are less frequent than in secondary and tertiary

referral centers. The major limitation is the uncontrolled setting. We do not know whether a drug

was discontinued after having been given at the highest tolerated dose. Our aim, however, was to

explore treatment changes as performed in clinical practice, where the selected daily dose generally

represents a compromise between seizure control, adverse events, and individual preference.

Another limitation is the time frame during which we started the follow-up. To include people with

newly diagnosed epilepsy in the cohort with reasonable follow-up, we started the observation at a

time in which mainly older AEDs were available. We are thus uncertain whether our findings apply

to cohorts starting treatment with a new AED. In keeping with our findings, however, there is no

evidence from more recent reports34-41 that new AEDs have advantages over older compounds.

Thus, we do not expect significant differences in other therapeutic contexts. Furthermore, the

cumulative time-dependent probability of withdrawal of AEDs other than carbamazepine,

phenobarbital and valproate was not assessed because of small numbers. The limited sample size

can also explain some non-significant correlations between demographic and clinical variables and

drug withdrawal. Lastly, we did our best to verify whether the indication for each drug was

appropriate. However, we cannot entirely exclude that drug failure was due to incorrect use of a

given drug in a given individual.

14. Giussani

In conclusion, the majority of people with epilepsy living in a community and followed for a

prolonged period of time remain treated with the first assigned drug. Seizure remission is the main

reason for drug discontinuation, followed by lack of efficacy and adverse events. Withdrawal of the

first drug for ineffectiveness and for adverse events tends to increase by AED order, while

decreasing for terminal remission. Withdrawal of the first AED for ineffectiveness can be predicted

by age at diagnosis while withdrawal of the second drug is predicted by seizure type, and reflects

the predominance of more severe epilepsy syndromes in younger individuals. These findings can

help the practicing physician to predict the response to the assigned treatment at diagnosis and when

a treatment change is required.

KEY POINTS

1. The initial antiepileptic drug (AED) given was retained in the majority of cases and only 10

percent of people started treatment with a new AED.

2. Reported reasons for AEDs withdrawal were, in decreasing order, terminal remission,

ineffectiveness and adverse events.

3. Discontinuation of a drug for terminal remission decreased with AED order; withdrawal for

ineffectiveness and adverse events increased.

4. Reasons for withdrawal varied with individuals’ age, sex, disease characteristics, and among

drugs.

5. The majority of people with epilepsy can be easily controlled with any of the available

compounds even after long follow-up periods.

ACKNOWLEDGEMENTS

The authors thank Gail Bell for technical editing. JWS is based at UCLH/UCL Comprehensive

Biomedical Research Centre, which receives a proportion of funding from the UK Department of

15. Giussani

Health’s National Institute for Health Research Biomedical Research Centres funding scheme. He

receives support from the Dr Marvin Weil Epilepsy Research Fund and UK Epilepsy Society.

The study was supported by an unrestricted educational grant from UCB-Pharma.

ETHICAL PUBLICATION

We confirm that we have read the Journal position on issue involved in ethical publication and

affirm that this report is consistent with those guidelines.

DISCLOSURE OF CONFLICTS OF INTEREST

EBe serves on the editorial boards of Amyotrophic Lateral Sclerosis, Clinical Neurology &

Neurosurgery, and Neuroepidemiology; has been an associate editor of Epilepsia; has received

money for board membership from VIROPHARMA and EISAI; has received funding for travel and

speaker honoraria from UCB-Pharma, Sanofi-Aventis, GSK; has received funding for educational

presentations from GSK; reports grants from the Italian Drug Agency and from the Italian Ministry

of Health. JWS has received research funding from Eisai, GSK and UCB, personal fees from Eisai,

UCB Lundbeck and Teva, outside the submitted work.

The remaining authors have nothing to disclose.

16. Giussani

References

1) Annegers JF, Hauser WA, Elverback LR. Remission of seizures and relapse in patients with

epilepsy. Epilepsia 1979;20:729-737.

2) Sillanpää M, Schmidt D. Natural history of treated childhood-onset epilepsy: prospective, long-

term population-based study. Brain 2006;129:617-624.

3) Kwan P, Brodie MJ. Effectiveness of first antiepileptic drug. Epilepsia 2001;42:1255-1260.

4) Brodie MJ, Barry SJ, Bamagous GA, Norrie JD, Kwan P. Patterns of treatment response in

newly diagnosed epilepsy. Neurology 2012;78:1548-1554.

5) Berg AT, Levy SR, Testa FM, D'Souza R. Remission of epilepsy after two drug failures in

children: a prospective study. Ann Neurol 2009;65:510-519.

6) Berg AT, Testa FM, Levy SR, Shinnar S, DiMario F, Smith S. How long does it take for

epilepsy to become intractable? A prospective investigation. Ann Neurol 2006;60:73-79.

7) Schiller Y. Seizure relapse and development of drug resistance following long-term seizure

remission. Arch Neurol 2009;66:1233-1239.

8) Kwan P, Arzimanoglou A, Berg AT, et al. Definition of drug resistant epilepsy: consensus

proposal by the ad hoc Task Force of the ILAE Commission on Therapeutic Strategies. Epilepsia

2010;51:1069-1077.

17. Giussani

9) Luciano AL, Shorvon SD. Results of treatment changes in patients with apparently drug-

resistant chronic epilepsy. Ann Neurol 2007;62:375-381.

10) Callaghan BC, Anand K, Hesdorffer D, Hauser WA, French JA. Likelihood of seizure

remission in an adult population with refractory epilepsy. Ann Neurol 2007;62:382-389.

11) Neligan A, Bell GS, Elsayed M, Sander JW, Shorvon SD. Treatment changes in a cohort of

people with apparently drug-resistant epilepsy: an extended follow-up. J Neurol Neurosurg

Psychiatry 2012;83:810-813.

12) Lhatoo SD, Sander JW, Shorvon SD. The dynamics of drug treatment in epilepsy: an

observational study in an unselected population based cohort with newly diagnosed epilepsy

followed up prospectively over 11-14 years. J Neurol Neurosurg Psychiatry 2001;71:632-637.

13) Wang WZ, Wu JZ, Ma GY, et al. Efficacy assessment of phenobarbital in epilepsy: a large

community-based intervention trial in rural China. Lancet Neurol 2006;5:46-52.

14) Huang DH, Zheng JO, Chen J, Yu L. Treatment gaps of epilepsy and retention rates of sodium

valproate in rural Guangxi, China. Genet Mol Res 2014;13:6202-6212.

15) Giussani G, Canelli V, Bianchi E, et al. A population-based study of active and drug-resistant

epilepsies in Northern Italy. Epilepsy Behav 2016;55:30-37.

16) Giussani G, Canelli V, Bianchi E, et al. Long-term prognosis of epilepsy, prognostic patterns

and drug resistance. A population-based study. Eur J Neurol 2016;23:1218-1227.

18. Giussani

17) Commission on Epidemiology and Prognosis, International League Against Epilepsy.

Guidelines for epidemiologic studies on epilepsy. Epilepsia 1993;34:592-596.

18) Commission on Classification and Terminology of the International League Against Epilepsy.

Proposal for revised clinical and electrographic classification of epileptic seizures. Epilepsia

1981;22:489-501.

19) Commission on Classification and Terminology of the International League Against Epilepsy.

Proposal for revised clinical classification of epilepsies and epileptic syndromes. Epilepsia

1989;30:389-399.

20) Berg AT, Berkovic SF, Brodie MJ, et al. Revised terminology and concepts for organization of

seizures and epilepsies: report of the ILAE Commission on Classification and Terminology, 2005-

2009. Epilepsia 2010;51:676-685.

21) Gray R. A class of k-sample tests for comparing the cumulative incidence of a competing risk.

The Ann Statistics 1988;16:1141–1154.

22) Dlugos DJ, Sammel MD, Strom BL, Farrar JT. Response to first drug trial predicts outcome in

childhood temporal lobe epilepsy. Neurology 2001;57:2259-2264.

23) Wirrell E, Camfield C, Camfield P, Dooley J. Prognostic significance of failure of the initial

antiepileptic drug in children with absence epilepsy. Epilepsia 2001;42:760-763.

24) Zhang Y, Yu N, Su L, Di Q. A prospective cohort study of prognosis for newly diagnoses

epilepsy in east China. BMC Neurology 2013;13:116.

19. Giussani

25) Beydoun A, Sawaya MT, Alam E, Hmaimess G, Ezzeddine K, Younes F. Treatment registry in

focal epilepsy (TRIP): multicenter observational study in Lebanon. Seizure 2015;27:54-59.

26) Bourgeois FT, Olson KL, Poduri A, Mandl KD. Comparison of Drug Utilization Patterns in

Observational Data: Antiepileptic Drugs in Pediatric Patients. Paediatr Drugs 2015;17(5):401-410.

27) Arif H, Buchsbaum R, Pierro J, et al. Comparative effectiveness of 10 antiepileptic drugs in

older adults with epilepsy. Arch Neurol 2010; 67: 408-415.

28) Chung S, Wang N, Hank N. Comparative retention rates and long-term tolerability of new

antiepileptic drugs. Seizure 2007;16:296-304.

29) Peltola J, Peltola M, Auvinen A, Raitanen J, Fallah M, Keränen T. Retention rates of new

antiepileptic drugs in localization-related epilepsy: a single-center study. Acta Neurol Scand

2009;119:55-60.

30) Bootsma HP, Ricker L, Hekster YA, et al. The impact of side effects on long-term retention in

three new antiepileptic drugs. Seizure 2009;18:327-331.

31) Kuba R, Novotná I, Brázdil M, et al. Long-term levetiracetam treatment in patients with

epilepsy: 3-year follow up. Acta Neurol Scand 2010;121:83-88.

32) Lhatoo SD, Wong IC, Sander JW. Prognostic factors affecting long-term retention of

topiramate in patients with chronic epilepsy. Epilepsia 2000;41:338-341.

20. Giussani

33) Hufnagel A, Kowalik A, Rettig K, Schreiner A, Schäuble B; TOP-GER-11 Investigators. Long-

term assessment of topiramate for epilepsy: an open-label, single-arm, multicentre, prospective

study in a naturalistic setting. Clin Drug Investig 2011;31:779-790.

34) Wilby J, Kainth A, Hawkins N, et al. Clinical effectiveness, tolerability and cost-effectiveness

of newer drugs for epilepsy in adults: a systematic review and economic evaluation. Health Technol

Assess 2005;9:1-157, iii-iv.

35) Marson AG, Al-Kharusi AM, Alwaidh M, et al. The SANAD study of effectiveness of

valproate, lamotrigine, or topiramate for generalised and unclassifiable epilepsy: an unblinded

randomised controlled trial. Lancet 2007;369:1016-1026.

36) Lee SK. Old versus New: Why Do We Need New Antiepileptic Drugs? J Epilepsy Res

2014;4:39-44.

37) Weijenberg A, Offringa M, Brouwer OF, Callenbach PM. RCTs with new antiepileptic drugs

in children: a systematic review of monotherapy studies and their methodology. Epilepsy Res

2010;91:1-9.

38) Tomson T. Drug selection for the newly diagnosed patient: when is a new generation

antiepileptic drug indicated? J Neurol. 2004;251:1043-1049.

39) Perucca E. Marketed new antiepileptic drugs: are they better than old-generation agents? Ther

Drug Monit 2002;24:74-80.

21. Giussani

40) Rowan AJ1, Ramsay RE, Collins JF, et al. New onset geriatric epilepsy: a randomized study of

gabapentin, lamotrigine, and carbamazepine. Neurology 2005;64:1868-1873.

41) Steinhoff BJ, Hirsch E, Mutani R, Nakken KO. The ideal characteristics of antiepileptic

therapy: an overview of old and new AEDs. Acta Neurol Scand 2003;107:87-95.

22. Giussani

Table 1. General characteristics of the sample (n=747)

Variable Category N %

Gender W 366 49.0

M 381 51.0

Family history of Yes 111 14.9

Seizures* No 584 78.2

Unknown 52 6.9

Seizures Partial 460 61.6

Generalized 260 34.8

Unclassifiable 27 3.6

Syndrome GC/GS 74 9.9

GI 154 20.6

PC 167 22.4

PI 54 7.2

PS 265 35.5

Undetermined 26 0.9

Special 7 3.5

Age at diagnosis <15y 318 42.7

15-64y 353 47.4

65+y 74 9.9

Missing 2

Disease duration <1y 666 89.5

at diagnosis ≥1 y 78 10.5

Missing 3

Number of AEDs 0 16 2.11 393 52.62 199 26.63 74 9.9

4+ 65 8.7

Legend: W women; M Men; GC/GS Generalized Cryptogenic/Generalized Symptomatic;GI Generalized Idiopathic; PC Partial Cryptogenic;PI Partial Idiopathic; PS Partial Symptomatic; y Years; AED Antiepileptic drug.*Any type of seizures in all known relatives.

23. Giussani

Table 2. Administration frequency of first, second and third drug by active principle

First drug Second drug Third drugFourth to ninth

drug

N % N % N % N

BSC 18 2.5 7 2.1 1 0.7 1

CBZ 203 27.8 60 17.8 18 12.9 4

CLB 2 0.3 22 6.5 12 8.6 7

CNP 2 0.3 31 9.2 8 5.8 16

ESM 9 1.2 4 1.2 4 2.9 2

GBP 2 0.3 6 1.8 4 2.9 6

LEV 13 1.8 32 9.5 18 12.9 29

LTG 4 0.5 32 9.5 13 9.4 15

OXC 31 4.2 16 4.7 5 3.6 10

PB 197 26.9 39 11.5 9 6.5 3

PHT 27 3.7 20 5.9 1 0.7 1

PGB 1 0.1 3 0.9 6 4.3 4

PRM 0 0.0 3 0.9 8 5.8 3

TGB 1 0.1 0 0.0 0 0.0 2

TPM 4 0.5 13 3.8 14 10.1 10

VGB 3 0.4 4 1.2 5 3.6 2

VPA 211 28.9 44 13.0 12 8.6 7

VPM 3 0.4 2 0.6 1 0.7 1

ZNS 0 0.0 0 0.0 0 0.0 3

Total Npatients*

731 338 139 65

Old AEDs 672 91.9 232 68.6 74 53.2 45NewAEDs

59 8.1 106 31.4 65 46.8 81

AED Antiepileptic drugs, BSC Barbexaclone, CBZ Carbamazepine, CLB Clobazam,

CNP Clonazepam, ESM Ethosuximide, GBP Gabapentin, LEV Levetiracetam, LTG Lamotrigine,

OXC Oxcarbazepine, PB Phenobarbital, PHT Phenytoin, PGB Pregabalin, PRM Primidone,

TGB Tiagabine, TPM Topiramate, VGB Vigabatrin, VPA Valproate, VPM Valpromide,

ZNS Zonisamide.

*16 did not start drugs.

24. Giussani

Table 3. Reasons of drug withdrawal by order

RankingPatientstreated

Ineffectiveness

Adverseevents

Terminalremission

Other*Never

withdrawn

N N % N % N % N % N %

First 731** 83 11.3 20 2.7 117 16.0 45 6.2 460 62.9

Second 338** 45 13.3 11 3.2 24 7.1 20 5.9 232 68.6

Third 139 33 23.7 6 4.3 4 2.9 6 4.3 90 64.7

Fourth 65 22 33.8 3 4.6 0 0.0 2 3.1 38 58.5

Fifth 31 11 35.5 2 6.5 2 6.5 0 0.0 16 51.6

Sixth 18 6 33.3 0 0.0 1 5.6 0 0.0 11 61.1

Seventh 7 3 42.9 0 0.0 0 0.0 0 0.0 4 57.1

Eighth 3 0 0.0 2 66.7 0 0.0 0 0.0 1 33.3

Ninth 2 0 0.0 1 50.0 0 0.0 0 0.0 1 50.0

*Death, drug out of production, pregnancy, own volition.

** Missing information about AED withdrawal in six.

25. Giussani

Table 4. Frequency of reasons of drug withdrawal by active principle

Patients

treatedIneffectiveness Adverse events

Terminal

remissionOther*

Never

withdrawn

N N % N % N % N % N %

BSC 27 6 22.2 1 3.7 4 14.8 9 33.3 7 25.9

CBZ 285 31 10.9 16 5.6 29 10.2 15 5.3 194 68.1

CLB 43 12 27.9 0 0.0 2 4.7 2 4.7 27 62.8

CNP 57 6 10.5 1 1.8 1 1.8 2 3.5 47 82.5

ESM 19 4 21.1 0 0.0 6 31.6 0 0.0 9 47.4

GBP 18 12 66.7 0 0.0 0 0.0 0 0.0 6 33.3

LEV 92 12 13.0 5 5.4 3 3.3 0 0.0 72 78.3

LTG 64 15 23.4 1 1.6 2 3.1 4 6.3 42 65.6

OXC 62 10 16.1 3 4.8 5 8.1 3 4.8 41 66.1

PB 248 34 13.7 7 2.8 34 13.7 23 9.3 150 60.5

PHT 49 16 32.7 3 6.1 2 4.1 4 8.2 24 49.0

PGB 14 1 7.1 0 0.0 0 0.0 2 14.3 11 78.6

PRM 14 2 14.3 0 0.0 1 7.1 1 7.1 10 71.4

TGB 3 2 66.7 1 33.3 0 0.0 0 0.0 0 0.0

TPM 41 11 26.8 4 9.8 4 9.8 1 2.4 21 51.2

VGB 14 7 50.0 0 0.0 2 14.3 0 0.0 5 35.7

VPA 274 27 9.9 3 1.1 55 20.1 8 2.9 181 66.1

VPM 7 0 0.0 0 0.0 1 14.3 1 14.3 5 71.4

ZNS 3 0 0.0 0 0.0 0 0.0 0 0.0 3 100.0

Old AEDs 1023 138 13.5 31 3.0 135 13.2 66 6.5 654 63.9

New AEDs 311 70 22.2 14 4.5 16 5.1 10 3.2 201 64.6AED Antiepileptic drug, BSC Barbexaclone, CBZ Carbamazepine, CLB Clobazam, CNP Clonazepam, ESM Ethosuximide,

GBP Gabapentin, LEV Levetiracetam, LTG Lamotrigine, OXC Oxcarbazepine, PB Phenobarbital, PHT Phenytoin,

PRG Pregabalin, PRM Primidone, TGB Tiagabine, TPM Topiramate, VGB Vigabatrin, VPA Valproate, VPM Valpromide,

ZNS Zonisamide. * Death, drug out of production, pregnancy, own volition withdrawal.

26. Giussani

Figure 1. Cumulative incidence functions for withdrawal of carbamazepine, phenobarbital

and valproate, for ineffectiveness (A) and terminal remission (B)